MC 5.5.3

I have 2 audio clips on the same track with a long cross fade between them. The second clip has a single Auto Gain keyframe and the level is reduced to the bottom.

When I play this I would expect to hear the first clip fading away to nothing (because the second clip is effectively muted by the gain) but what I hear is the first clip fading into the second clip until halfway through the fade and then falling away to silence.

Is this normal? I don't remember this behaviour before.

Where in relation to the transition is the audio keyframe? FWIW you can't have just one audio keyframe (well you can, but it will be ineffective), you need one at the start of the audio level adjustment and one at the end. If you trim the edit way over to the right are you sure that the level of the outgoing clip is in fact keyframed down beyond the edit point?

James, AutoGain comes after ClipGain. Dissolves will interpolate between AutoGain keyframes as well.

So yes, what you are hearing is a dissolve between clip 1&2, and on top of that, a transition in AutoGain from 0dB to -Infinity.

If you want to mute the second clip, use Clip Gain to lower it to -Infinity.

Here is a screengrab.

If you play this you will hear the second clip during the crossfade.

If you take away the crossfade then you will not hear the second clip.

It is possible to have one Auto Gain (Volume in MC6) keyframe. It will set the gain for the entire clip. It is a much easier way to manipulate levels on clips than using the Clip Gain in the mixer.

Hi Job,

I was afraid you might say that it is behaving as it is designed to.

I would argue that it is therefore not a very usefull arangement. To be more usefull it should behave like clip gain>auto gain>cross fade. Why would the current order be a benefit?

In my example, let's say I wanted to crossfade to the second clip at a low level (not -infinity) and then after the fade have the clip ramp up to full volume. If I had to use clip gain to to achieve the initial low level then I wouldn't be able to do this.

I would love to hear other opinions on this. How many editors would expect to be hearing the second clip in the arrangement that I posted? How would the same arrangement behave in Pro-Tools?

EDIT:

Wouldn't it be better to not have keyframes be interpolated between clips? I'm finding it hard to understand why it is a benefit.

I just did a test with my Final Cut friend.

I asked her to set up the exact same scenario. I asked her what she expected to hear and she said that she did not expect to hear the second clip and in fact you don't hear the second clip.

+1 for FCP in my opinion.

The mystery deepens. I just tried the exact same experiment on V 3.1.3 and it behaves just like FCP. Why did they change it?

I have an editor now who is very upset and screaming blue murder at Avid for making his old workflow obsolete.

You could be very right that this behaviour used to be diffent, and I vaguely recall running it to it once or twice, last year. Not sure when it changed, but I'd suspect 6.x. Also not sure if this was intentional. I doubt it has anything to do with ProTools, as that has sort of a separate layer for automation per track.

The only reason I can think of is the scenario where you have two pieces of music, both with Automation Gain set, and you crossfade between them. You would not want AutoGain to jump up or down in the middle of that cross.

The absolute downside (besides pissing off your editor) is that it breaks compatibility with older sequences.

It has been changed by version 5.5 at the latest.

Unless there is a big workflow reason why the new way is better then I would strongly push for an option to have it behave the old way.

There are many times when I want to use keyframes and a crossfade at the same time and currently the results are not intuitive or predictable and it seems the only way to avoid this problem is to use more tracks (Not easy when you are already on 16).

I thought maybe it was changed to match the behaviour of Pro-Tools. This would make sense but there should be an option for backwards compatability.

Any chance you can make some enquiries with the devs?

I will see what I can find out, but I'm on holiday in France, so it may take me until next week to undertake proper action. Please bump the thread if I forget about it...

James,

Turns out I was all wrong about this in my first response, it's a known bug. Our own Super-Marianna just informed me that the bug is that Audio Keyframing is ignored during dissolves, and that this is on fact a known bug, listed under UDevC00172587. So it's known, and it's not how it's supposed to be.

I'm glad to see that it has been recognised as a bug. It seems a little more than just keyframes being ignored however.

Take a look at this screenshot comparing the two version and how the mixdowns behave.

Bump. Any news on this? I'm on 6.5 and the behavior is still the same. 

Moreover, there seems to be a difference in how MC handles dissolves over audio keyframes depending on whether you've rendered your dissolve or not. In the OP's original case the behavior is exactly as he describes regardless of whether you render, but sometimes I find that dissolving between two clips with keyframed audio (and not mute keyframes) will play correctly only when rendered, despite the usual green dot indicating that the dissolve will play in real-time. 

In that case the real-time dissolve will use the clip gain only of the B-side clip and disregard the auto gain up until the dissolve is over, when it will then jump to the auto gain volume. Only when you render is the auto gain taken into account by the dissolve.

Hope that was clear! :)

Hi Evan,

I'm still on 5.5 but I am shocked if you say this is not fixed in 6.5 yet.

This is a MAJOR bug if a simple audio crossfade behaves in the stupid way as I illustrated.Every edit has dozens of such fades.

This bug means any audio work that was done pre 5 will now sound wrong in 5 and above.

come on Avid!! Sort out this basic bug please!!

I'm bumping my own post as this bug is still in 6.5. I can't believe more people aren't concerned with it.

To explain it further try the following experiment.

Take an audio clip (try some dense music) in the source window and apply a -10db gain to it

Place the clip of audio in the timeline and add an edit in the middle

Add a cross fade of 2 seconds to the edit

(if you play this you hear just the normal audio as you would expect)

Now add one Auto Gain (Volume) key frame to the second part of the clip and make sure it is left at 0db.

Now when you play it you will hear the audio dip during the cross fade.